Not JUCE related, but the first time I came across it is in the JUCE framework and I know Jules etc. are big on doing things the proper way, so curious to know the musings/advice from the JUCErs here.
I’m really incompetent at programming, in fact, mid way through crunching Bruce Eckel’s ‘Thinking in C++’ I decided to go through another book (Teach yourself C++ in one hour a day) before I could even finish the first volume of Thinking. This is how amateur I am.
Anyway, for dumb dumbs like me, the auto type specifier seems to be more pain than it’s worth, unless I’m missing something? I like knowing easily what type something is, so what is the benefit to using it, and why is it used so often in JUCE?
Another argument, your code might not even mention the type:
auto value = getFoo();
If I later refactor getFoo to return double instead of float, my code still does the expected thing, just in a higher precision.
What you should care about, when you use auto, is to add const if applicable and * (it’s a pointer) or & (it’s a reference, not take the address of).
So generally auto is a case of DRY and avoiding unnecessary implicit casts.
auto also has the advantage that it’s impossible to leave it uninitialised.
int x; // maybe 0 depending on context, but probably indeterminate
auto x = 0; // definitely 0
auto x = int { 0 }; // definitely 0 and obviously an int
